Potential Objects Of Interest Research Articles

Autonomous Underwater Environment Perceiving and Modeling: An Experimental Campaign With FeelHippo AUV for Forward Looking Sonar-Based Automatic Target Recognition and Data Association

Seabed inspection is one of the most sought-after applications for autonomous underwater vehicles (AUVs). Acoustical sensors, such as side-scan sonars and forward-looking sonars (FLSs), are commonly favored over optical cameras to carry out such a task. Indeed, sonars are not influenced by illumination conditions and can provide high-range data. However, due to the lack of features and low resolution, acoustical images are often hard to interpret with conventional automatic techniques, forcing human operators to analyze thousands of collected images to identify the so-called objects of potential interest (OPIs). In this article, we report the development of an automatic target recognition (ATR) methodology to identify and localize OPIs in FLS imagery. Such detections have been then exploited to realize a virtual world model with the probabilistic multiple hypothesis anchoring data association and model tracking algorithm. Distinct models of convolutional neural networks have been trained with a data set acquired in May 2019 at the Naval Support and Experimentation Centre (Centro di Supporto e Sperimentazione Navale—CSSN) basin in La Spezia, Italy. The ATR strategy has been successfully validated offline with the data gathered in October 2019 in the same site where the seabed targets were replaced and relocated. As regards the world modeling technique, it has been preliminarily tested on a simulated scenario built upon unmanned underwater vehicle Simulator. Finally, both the ATR and world modeling systems were on-field tested in October 2020 at the CSSN basin in a multivehicle architecture by employing an acoustical channel between FeelHippo AUV and an autonomous moving buoy.

Detection and classification of man-made objects for the autonomy of underwater robots

Recent developments in marine technologies allow underwater vehicles to perform survey missions for data collection in an automatic way. The scientific community is now focusing on endowing these vehicles with strong perception capabilities, aiming at full autonomy and decision-making skills. Such abilities would bring benefits to a wide range of field applications, e.g. Inspection and Maintenance (I&M) of man-made structures, port security, and marine rescue. Indeed, most of these tasks are currently carried out employing remotely operated vehicles, making the presence of humans in water necessary. Projects like Metrological Evaluation and Testing of Robots in International CompetitionS (METRICS), funded by the European Commission, are promoting research on this field by organising events such as the Robotics for Asset Maintenance and Inspection (RAMI) competition. In particular, this competition requires participants to develop perception techniques capable of identifying a set of specific targets. Within such context, this paper presents an algorithm able to detect and classify Objects of Potential Interest (OPIs) in underwater camera images. First, the proposed solution compensates for the quality degradation of underwater images by applying color enhancement and restoration procedures. Then, it exploits deep-learning techniques, as well as color and shape based methods, to recognize and correctly label the predefined OPIs. Preliminary results of the implemented neural network using restored images are provided, and a mean Average Precision (mAP) of about 92% was achieved on the dataset provided to the RAMI competition participating teams by the NATO Science and Technology Organization Centre for Maritime Research and Experimentation (STO CMRE).

A Very High Energy γ-Ray Survey toward the Cygnus Region of the Galaxy

Abstract We present results from deep observations toward the Cygnus region using 300 hr of very high energy (VHE) γ-ray data taken with the VERITAS Cerenkov telescope array and over 7 yr of high-energy γ-ray data taken with the Fermi satellite at an energy above 1 GeV. As the brightest region of diffuse γ-ray emission in the northern sky, the Cygnus region provides a promising area to probe the origins of cosmic rays. We report the identification of a potential Fermi-LAT counterpart to VER J2031+415 (TeV J2032+4130) and resolve the extended VHE source VER J2019+368 into two source candidates (VER J2018+367* and VER J2020+368*) and characterize their energy spectra. The Fermi-LAT morphology of 3FGL J2021.0+4031e (the Gamma Cygni supernova remnant) was examined, and a region of enhanced emission coincident with VER J2019+407 was identified and jointly fit with the VERITAS data. By modeling 3FGL J2015.6+3709 as two sources, one located at the location of the pulsar wind nebula CTB 87 and one at the quasar QSO J2015+371, a continuous spectrum from 1 GeV to 10 TeV was extracted for VER J2016+371 (CTB 87). An additional 71 locations coincident with Fermi-LAT sources and other potential objects of interest were tested for VHE γ-ray emission, with no emission detected and upper limits on the differential flux placed at an average of 2.3% of the Crab Nebula flux. We interpret these observations in a multiwavelength context and present the most detailed γ-ray view of the region to date.

Automated data processing architecture for the Gemini Planet Imager Exoplanet Survey

The Gemini Planet Imager Exoplanet Survey (GPIES) is a multi-year direct imaging survey of 600 stars to discover and characterize young Jovian exoplanets and their environments. We have developed an automated data architecture to process and index all data related to the survey uniformly. An automated and flexible data processing framework, which we term the Data Cruncher, combines multiple data reduction pipelines together to process all spectroscopic, polarimetric, and calibration data taken with GPIES. With no human intervention, fully reduced and calibrated data products are available less than an hour after the data are taken to expedite follow-up on potential objects of interest. The Data Cruncher can run on a supercomputer to reprocess all GPIES data in a single day as improvements are made to our data reduction pipelines. A backend MySQL database indexes all files, which are synced to the cloud, and a front-end web server allows for easy browsing of all files associated with GPIES. To help observers, quicklook displays show reduced data as they are processed in real-time, and chatbots on Slack post observing information as well as reduced data products. Together, the GPIES automated data processing architecture reduces our workload, provides real-time data reduction, optimizes our observing strategy, and maintains a homogeneously reduced dataset to study planet occurrence and instrument performance.

Evaluation of Attention‐Guiding Video Visualization

AbstractWe investigate four different variants of attention‐guiding video visualization techniques that aim to help users distribute their attention equally among potential objects of interest: bounding box visualization, force‐directed visualization, top‐down visualization, grid visualization. Objects of interest are highlighted by rectangular shapes and then we concentrate on the manipulation of color, motion, and size. We conducted a controlled laboratory user study (n=25) to compare the four visualization techniques and the unmodified video material as baseline. We evaluated task performance and distribution of attention in a search task. These two properties become especially important when video material with numerous objects has to be observed. The distribution of attention was measured by eye tracking. Our results show that a more even distribution of attention between the objects can be achieved by attention‐guiding visualization, compared to unmodified video. Many participants feel more comfortable when they look at bounding boxes and the grid, but improvements in search task performance could not be confirmed.

Saccade target selection in the superior colliculus during a visual search task.

Because real-world scenes typically contain many different potential objects of interest, selecting one goal from many is clearly a fundamental problem faced by the saccadic system. We recorded from visual, movement, and visuo-movement (VM) neurons in the superior colliculus (SC) of monkeys performing a reaction-time visual-search task requiring them to make saccades to an odd-colored target presented with distractors. First, we compared the responses of SC neurons in search with their responses when a single target was presented without distractors (single-stimulus task). Consistent with earlier reports, initial visual activity was smaller in search than in the single-stimulus task, while movement-related activity in the two tasks was comparable. Further experiments showed that much of the reduction in the initial visual response during search was due to lateral inhibition, although a top-down task-related component was also evident. Although the initial visual activity did not discriminate the target from the distractors, some neurons showed a biphasic pattern of visual activity. In VM burst neurons, the second phase of this activity was significantly larger when the target, rather than a distractor, was in the response field. We traced the time course of target/distractor discrimination using receiver operating characteristic (ROC) analysis and found that VM burst neurons, VM prelude neurons, and pure movement neurons discriminated the target from distractors before saccade onset but that phasic and tonic pure visual neurons did not. We also examined the relationship between target/distractor discrimination time and saccade latency. Discrimination in VM burst neurons having a biphasic pattern of visual activity and in many VM prelude neurons occurred after a consistent delay that did not depend on saccade latency, suggesting that these neurons are involved in target selection as well as movement initiation. In contrast, VM burst neurons lacking a biphasic pattern of visual activity, pure movement neurons, and a subset of VM prelude neurons discriminated the target at a time that was well correlated with saccade latency, suggesting that this latter group of neurons is involved in triggering movement execution but not in target selection. Thus a mix of signals likely related to target selection and movement initiation co-exists in different groups of SC neurons. This suggests that certain types of SC neurons participate in the target selection process and that the SC as a whole represents a gateway for target selection signals to be converted into a saccadic command.